In New York City HIV testing of street addicts shows 50% already rare seropositive and are thus capable of transmitting the virus. Jails in large cities have high concentrations (about 25%) of untreated, intravenous drug-abusing street addicts who, at release, return to AIDS risk behaviors such as drug injection, needle-sharing, and unprotected sexual activity. IVDU females are a main source of perinatal and (potentially) heterosexual HIV transmission. Other jail inmates are cocaine/crack-using adolescents at high risk of progressing to intravenous drug use, and then contracting/transmitting HIV through needle use and sex. These groups are very difficult to reach and work with in the street milieu. The study proposes to implement and evaluate a comprehensive AIDS prevention program for three target groups--untreated male and female IVs and cocaine/crack-using male adolescents--while they are temporarily accessible in jail. A total of 720 inmates (240 in each target group) will be randomly assigned by jail ward to three study conditions: (1) enhanced in-jail AIDS prevention services and post-release community counseling (N = 300); (2) in-jail AIDS prevention services only (N = 210); and (3) standard AIDS orientation as a control (N = 210). The intensive in-jail services will consist of three components: (1) facilitated peer support groups using a cognitive-behavioral method, Problem-Solving Training (PST); (2) informal outreach with subjects on the jail wards; and (3) provision of an AIDS "survival package" (with bleach, condoms, etc.) at release. The post-release community counseling will reinforce the in-jail learning, help subjects deal with transition crises, and facilitate linkages to community support networks. Interview data will be collected at baseline, jail exit and 6-month follow-up (plus urinalysis at follow-up). Key outcome measures will include changes in: AIDS knowledge; attitudes/beliefs/intentions concerning AIDS risk reduction; IV drug use, needle-sharing and equipment cleaning practices; and relevant sex practices such as condom use. If shown effective this AIDS prevention model would be suitable for replication in large-city jail systems throughout the nation.